New 2ch pocket DSO+SG - Zeeweii DSO2512G

[Dosergen]

Greetings forum members. Could you tell me , if it's possible update to version 1.3.0C for the oscilloscope revision (HW 9.15 / SW 1.2.9C)?

[ptluis]

Greetings forum members. Could you tell me , if it's possible update to version 1.3.0C for the oscilloscope revision (HW 9.15 / SW 1.2.9C)?

v1.3.0C is the update version for v1.2.9C. I think new batch scopes come with the new version installed. unfortunately I don't have the v1.3.0C available. I suggest you to contact zeeweii and see if they send you the firmware.

[timschuerewegen]

Some interesting things about the firmware:
- it contains seemingly fully functional code to read/write from/to the microsd card (but it is not used because of a "use_sdcard" global variable that gets initialised to 0)
- images are saved to spi flash as 2048 byte raw data blobs (2x300 bytes for ch1, 2x300 bytes for ch2, and some other data)
- usb image transfer mode creates a new/empty ramdisk, reads the saved blobs from spi flash, renders them and creates bmp files on the ramdisk
- usb firmware update mode also creates a new/empty ramdisk
- there are 2x15000 bytes waveform buffers per channel, so 50 screens worth of waveform data

[RAPo]

How did you get access to the raw data blobs?

Some interesting things about the firmware:
- it contains seemingly fully functional code to read/write from/to the microsd card (but it is not used because of a "use_sdcard" global variable that gets initialised to 0)
- images are saved to spi flash as 2048 byte raw data blobs (2x300 bytes for ch1, 2x300 bytes for ch2, and some other data)
- usb image transfer mode creates a new/empty ramdisk, reads the saved blobs from spi flash, renders them and creates bmp files on the ramdisk
- usb firmware update mode also creates a new/empty ramdisk
- there are 2x15000 bytes waveform buffers per channel, so 50 screens worth of waveform data

[timschuerewegen]

I did not. I am reverse engineering the 1.2.8C firmware while I wait for my DSO2512G to arrive. My ultimate goal is to create a fully compilable 100% asm version of the firmware on my DSO2512G (hopefully the latest version) so that I can make modifications to the firmware without any restrictions and loss of functionality. The parts of the code that I want to modify will be converted to C but most of it can stay in pure asm like low level usb stuff, etc. and unknown stuff.

How did you get access to the raw data blobs?

[RAPo]

Ah so, thanks.

I did not. I am reverse engineering the 1.2.8C firmware while I wait for my DSO2512G to arrive. My ultimate goal is to create a fully compilable 100% asm version of the firmware on my DSO2512G (hopefully the latest version) so that I can make modifications to the firmware without any restrictions and loss of functionality. The parts of the code that I want to modify will be rewritten in C but most of it can stay in
pure asm like low level usb stuff, etc. and unknown stuff.

[ptluis]

I did not. I am reverse engineering the 1.2.8C firmware while I wait for my DSO2512G to arrive. My ultimate goal is to create a fully compilable 100% asm version of the firmware on my DSO2512G (hopefully the latest version) so that I can make modifications to the firmware without any restrictions and loss of functionality. The parts of the code that I want to modify will be converted to C but most of it can stay in pure asm like low level usb stuff, etc. and unknown stuff.

How did you get access to the raw data blobs?

If you bought your unit directly from zeeweii you probably get firmware 1.3.0C, from other sellers it's a kind of lottery.

[timschuerewegen]

My DSO2512G arrived. It has PCB V9.15 and FW V1.3.0C.

Hacking is fun:
https://youtu.be/3M0R_1HX7KE

[ptluis]

My DSO2512G arrived. It has PCB V9.15 and FW V1.3.0C.

Hacking is fun:
https://youtu.be/3M0R_1HX7KE

What are the main differences in code between 1.2.8C and 1.3.0C, did you manage to find out?

[timschuerewegen]

No,  I don't know the differences between V1.2.8C+ and V1.3.0C.

I just added .wav file export to my modified firmware, what else can I add/change

[Aldo22]

I just added .wav file export to my modified firmware, what else can I add/change

Cool! 
I also wanted to export wave forms to import them into the AWG.
But since I have no idea about firmware hacking, I just tried to read the information from the bitmap. The little software is not ready yet but the wave form looks better than the original! 

Will you share your firmware changes with us?

[timschuerewegen]

Sure. But you should make a backup of your SPI flash chip first in case something goes wrong. I desoldered my SPI flash chip and used a cheap USB BIOS programmer to dump it. I am also using it to reprogram the SPI flash chip to test my firmware changes. It should also be possible to read/write the SPI flash chip via USB "FEL" mode but only if you can insert a microSD card into the microSD slot, so you can enter USB FEL mode, which is impossible to do since the battery connector is in the way 

If anyone wants to discuss DSO2512G firmware hacking with me:
https://discord.gg/7UN88dhmKs

Will you share your firmware changes with us?

[timschuerewegen]

The first 300 bytes in the .wav blob are 300 Y1 values, the next 300 bytes are 300 Y2 values. Draw 300 vertical lines between Y1 and Y2 and you see the waveform.

I also wanted to export wave forms to import them into the AWG.
But since I have no idea about firmware hacking, I just tried to read the information from the bitmap. The little software is not ready yet but the wave form looks better than the original! 

[Aldo22]

The first 300 bytes in the .wav blob are 300 Y1 values, the next 300 bytes are 300 Y2 values. Draw 300 vertical lines between Y1 and Y2 and you see the waveform.

Great! 
But you don't get more resolution this way, right?
One can also get 300x200 px by scanning the bitmap, although it's a bit more complicated and possibly less reliable.

[timschuerewegen]

But you don't get more resolution this way, right?

I modified the firmware to save some fpga/waveform buffers to the SPI flash and one of the buffers, the one used to calculate all the measurements (pkpk, min, max, ...), contained 1200 bytes so I plotted them

[Aldo22]

But you don't get more resolution this way, right?

I modified the firmware to save some fpga/waveform buffers to the SPI flash and one of the buffers, the one used to calculate all the measurements (pkpk, min, max, ...), contained 1200

Not bad! 

But if the Y resolution remains the same, it's probably not extremely useful.
Do you manage to get to the whole buffer to see more of the timeline?
Like you can see in the zoom tool in the upper area?
That would perhaps be a real advantage
Sorry if I'm asking stupid questions. I have no idea what is possible this way.

[timschuerewegen]

Do you manage to get to the whole buffer to see more of the timeline?
Like you can see in the zoom tool in the upper area?
That would perhaps be a real advantage

In zoom mode the firmware requests/reads memory from the FPGA two times instead of one, using a different time base, so it can draw the upper and lower waveforms, so what you see on the LCD is probably all the data there is.

[timschuerewegen]

What are the main differences in code between 1.2.8C and 1.3.0C, did you manage to find out?

I did find a feature in V1.3.0C that does not exist in V1.2.8C+, a "real time" auto mode that can be activated by long pressing the AUTO key.

[timschuerewegen]

I made a one-page cheat sheet for myself if you're interested.

I found a "Gain calib" (V1.2.8C) and "Factory calib" (V1.2.8C+ and newer) mode, that can be actived from the "Meas" CH2 menu by long pressing DOWN.

I also found a reset option for some of the settings that can be activated from the "Set" menu by long pressing the AUTO key. All FW versions.

[Aldo22]

I did find a feature in V1.3.0C that does not exist in V1.2.8C+, a "real time" auto mode that can be activated by long pressing the AUTO key.

Yes, ptluis called it "continuous auto mode"
https://www.eevblog.com/forum/testgear/new-2ch-pocket-dsosg-sigpeak-dso2512g/msg4984915/#msg4984915

I found a "Gain calib" (V1.2.8C) and "Factory calib" (V1.2.8C+ and newer) mode, that can be actived from the "Meas" CH2 menu by long pressing DOWN.

Yes, i had already updated the cheat sheet. See "DSO2512g_plus_cheat.pdf"
https://www.eevblog.com/forum/testgear/new-2ch-pocket-dsosg-sigpeak-dso2512g/msg4722281/#msg4722281

[timschuerewegen]

Yes, i had already updated the cheat sheet.

The "long press down" one is not in your cheat sheet though, there are 3 calib modes:

Long press DOWN = "Factory calib" (all versions)
Long press V = "Calib:DC offset" (V1.2.8C+ and newer)
Long press mV = Calib:AC gain" (V1.2.8C+ and newer)

[Aldo22]

Yes, i had already updated the cheat sheet.

The "long press down" one is not in your cheat sheet though, there are 3 calib modes:

Long press DOWN = "Factory calib" (all versions)
Long press V = "Calib:DC offset" (V1.2.8C+ and newer)
Long press mV = Calib:AC gain" (V1.2.8C+ and newer)

OK, and what is "Factory calib"?
Is this the same as can be achieved from the "Aux" menu?

[timschuerewegen]

OK, and what is "Factory calib"?
Is this the same as can be achieved from the "Aux" menu?

No. I checked the code and it does exactly the same as "Calib:AC gain" (long press mV) but displays "Gain calib mode" instead of "Exit" when you leave that mode.

[Aldo22]

No. I checked the code and it does exactly the same as "Calib:AC gain" (long press mV) but displays "Gain calib mode" instead of "Exit" when you leave that mode.

Thank you.
I'm not going to update the cheat sheet because of this.
The cheat sheet is not necessarily comprehensive, but it serves to
- remember shortcuts for frequently used functions
- remember functions that only work with keyboard shortcuts.

And I'd like to keep it on one A4 page for printing, so there must be a good reason to change it. 

what else can I add/change

As I said before, I don't know what you can do, but you could fix bugs, for example. 

The DSO2512G is actually quite usable, but there is a problem in negative DC range.
Maybe this is already fixed in newer firmware, but have a look.

Try e.g. -3 Volt Offset DC.
On my device the grid position and the "mean" value do not match.
When the line on the grid is at -3 volts, the "mean" value shows e.g. -3.2 volts.
This looks like a software problem to me, otherwise both readings would be wrong to the same degree.

[timschuerewegen]

The DSO2512G is actually quite usable, but there is a problem in negative DC range.
Maybe this is already fixed in newer firmware, but have a look.

V1.3.0C has the same "bug". I have the formula that calculates the "Mean" value from the raw FPGA data so I will check if it is a bug or a feature